The following relates generally to signal processing, and more specifically to seamless listen-through for a wearable device.
A user may use a wearable device, and may wish to experience a listen-through feature. In some examples, when a user speaks (e.g., generates a self-voice signal), the user's voice may travel along two paths: an acoustic path and a bone conduction path. However, distortion patterns from external or background signals may be different than distortion patterns created by self-voice signals. Microphones picking up an input audio signal (e.g., including background noise and self-voice signals) may not seamlessly deal with the different types of signals. The different distortion patterns for different signals may result in a lack of natural sounding audio input when using a listen-through feature on the wearable device.